Breath analysis is emerging as a universal diagnostic
method for
clinical applications. The possibility of breath analysis is being
explored vigorously using different analytical techniques. We have
designed and assembled a multiwavelength UV photoacoustic spectroscopy
(PAS) sensor for the said application. To optimize laser wavelength
for sample excitation, photoacoustic signals from disease and normal
conditions are recorded with different laser excitations (213, 266,
355, and 532 nm) on exhaled breath samples. Principal component analysis
(PCA) of the PA signals has shown that 213, 266, and 355 nm laser
excitations are suitable for breath analysis, with reliable descriptive
statistics obtained for 266 nm laser. The study has, therefore, been
extended for breath samples collected from asthma, chronic obstructive
pulmonary disease (COPD), and normal subjects, using 266 nm laser
excitation. PCA of the PA data shows good classification among asthma,
COPD, and normal subjects. Match/No-match study performed with asthma,
COPD, and normal calibration set has demonstrated the potential of
using this method for diagnostic application. Sensitivity and specificity
are observed as 88 and 89%, respectively. The area under the curve
of the ROC curve is found to be 0.948, which justifies the diagnostic
capability of the device for lung diseases. The same samples were
studied using a commercial E-Nose, and the measurement outcome strongly
supports the PAS results.